CASPER in the Machine: Insights into Character Variety in LLM-Generated Stories

This study analyzes character variety in stories generated by large language models (LLMs) compared to human-written narratives, focusing on eight dimensions of character defined by narratology, such as stylization and wholeness. By categorizing characters in both sets of stories, the research addresses whether LLMs produce similar character types and the diversity of characters in their narratives. The findings provide insights into the capabilities and limitations of LLMs in creative writing, which is crucial for practitioners aiming to enhance narrative generation in AI applications.